|
|
| version 1.33.2.19.2.25, 2003/04/11 07:44:44 | version 1.51, 2005/04/08 10:40:18 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: Charset connection implementation. | Parser: Charset connection implementation. |
| Copyright(c) 2001-2003 ArtLebedev Group (http://www.artlebedev.com) | Copyright(c) 2001-2004 ArtLebedev Group (http://www.artlebedev.com) |
| Author: Alexander Petrosyan<paf@design.ru>(http://paf.design.ru) | Author: Alexander Petrosyan<paf@design.ru>(http://paf.design.ru) |
| */ | */ |
| static const char* IDENT_CHARSET_C="$Date$"; | static const char * const IDENT_CHARSET_C="$Date$"; |
| #include "pa_charset.h" | #include "pa_charset.h" |
| #include "pa_charsets.h" | #include "pa_charsets.h" |
| Line 14 static const char* IDENT_CHARSET_C="$Dat | Line 14 static const char* IDENT_CHARSET_C="$Dat |
| #include "libxml/encoding.h" | #include "libxml/encoding.h" |
| #endif | #endif |
| //#define PA_PATCHED_LIBXML_BACKWARD | |
| // globals | |
| Charset::UTF8CaseTable::Rec UTF8CaseToUpperRecords[]={ | |
| #include "utf8-to-upper.inc" | |
| }; | |
| Charset::UTF8CaseTable UTF8CaseToUpper={ | |
| sizeof(UTF8CaseToUpperRecords)/sizeof(Charset::UTF8CaseTable::Rec), | |
| UTF8CaseToUpperRecords}; | |
| Charset::UTF8CaseTable::Rec UTF8CaseToLowerRecords[]={ | |
| #include "utf8-to-lower.inc" | |
| }; | |
| Charset::UTF8CaseTable UTF8CaseToLower={ | |
| sizeof(UTF8CaseToLowerRecords)/sizeof(Charset::UTF8CaseTable::Rec), | |
| UTF8CaseToLowerRecords}; | |
| // helpers | // helpers |
| inline void prepare_case_tables(unsigned char *tables) { | inline void prepare_case_tables(unsigned char *tables) { |
| Line 68 static void element2case(unsigned char f | Line 86 static void element2case(unsigned char f |
| // methods | // methods |
| extern "C" unsigned char pcre_default_tables[]; // pcre/chartables.c | extern "C" unsigned char pcre_default_tables[]; // pcre/chartables.c |
| Charset::Charset(Request_charsets* charsets, const StringBody ANAME, const String* afile_spec): | Charset::Charset(Request_charsets* charsets, const String::Body ANAME, const String* afile_spec): |
| FNAME(ANAME), | FNAME(ANAME), |
| FNAME_CSTR(ANAME.cstrm()) { | FNAME_CSTR(ANAME.cstrm()) { |
| Line 101 void Charset::load_definition(Request_ch | Line 119 void Charset::load_definition(Request_ch |
| // charset | // charset |
| memset(&tables, 0, sizeof(tables)); | memset(&tables, 0, sizeof(tables)); |
| // strangly vital | |
| tables.toTable[tables.toTableSize].intCh=0; | |
| tables.toTable[tables.toTableSize].extCh=(XMLByte)0; | |
| tables.toTableSize++; | |
| // loading text | // loading text |
| char *data=file_read_text(charsets, afile_spec); | char *data=file_read_text(charsets, afile_spec); |
| Line 114 void Charset::load_definition(Request_ch | Line 128 void Charset::load_definition(Request_ch |
| // parse cells | // parse cells |
| char *row; | char *row; |
| while(row=getrow(&data)) { | while((row=getrow(&data))) { |
| // remove empty&comment lines | // remove empty&comment lines |
| if(!*row || *row=='#') | if(!*row || *row=='#') |
| continue; | continue; |
| Line 122 void Charset::load_definition(Request_ch | Line 136 void Charset::load_definition(Request_ch |
| // char white-space digit hex-digit letter word lowercase unicode1 unicode2 | // char white-space digit hex-digit letter word lowercase unicode1 unicode2 |
| unsigned int c=0; | unsigned int c=0; |
| char *cell; | char *cell; |
| for(int column=0; cell=lsplit(&row, '\t'); column++) { | for(int column=0; (cell=lsplit(&row, '\t')); column++) { |
| switch(column) { | switch(column) { |
| case 0: c=to_wchar_code(cell); break; | case 0: c=to_wchar_code(cell); break; |
| // pcre_tables | // pcre_tables |
| Line 161 void Charset::load_definition(Request_ch | Line 175 void Charset::load_definition(Request_ch |
| static int sort_cmp_Trans_rec_intCh(const void *a, const void *b) { | static int sort_cmp_Trans_rec_intCh(const void *a, const void *b) { |
| return | return |
| static_cast<const Charset_TransRec *>(a)->intCh- | static_cast<const Charset::Tables::Rec *>(a)->intCh- |
| static_cast<const Charset_TransRec *>(b)->intCh; | static_cast<const Charset::Tables::Rec *>(b)->intCh; |
| } | } |
| void Charset::sort_ToTable() { | void Charset::sort_ToTable() { |
| Line 176 void Charset::sort_ToTable() { | Line 190 void Charset::sort_ToTable() { |
| static XMLByte xlatOneTo(const XMLCh toXlat, | static XMLByte xlatOneTo(const XMLCh toXlat, |
| const Charset::Tables& tables, | const Charset::Tables& tables, |
| XMLByte not_found) { | XMLByte not_found) { |
| unsigned int lowOfs = 0; | int lo = 0; |
| unsigned int hiOfs = tables.toTableSize - 1; | int hi = tables.toTableSize - 1; |
| XMLByte curByte = 0; | while(lo<=hi) { |
| do { | |
| // Calc the mid point of the low and high offset. | // Calc the mid point of the low and high offset. |
| const unsigned int midOfs =((hiOfs - lowOfs) / 2)+lowOfs; | const unsigned int i = (lo + hi) / 2; |
| // If our test char is greater than the mid point char, then | XMLCh cur=tables.toTable[i].intCh; |
| // we move up to the upper half. Else we move to the lower | if(toXlat==cur) |
| // half. If its equal, then its our guy. | return tables.toTable[i].extCh; |
| if(toXlat>tables.toTable[midOfs].intCh) | if(toXlat>cur) |
| lowOfs = midOfs; | lo = i+1; |
| else if(toXlat<tables.toTable[midOfs].intCh) | |
| hiOfs = midOfs; | |
| else | else |
| return tables.toTable[midOfs].extCh; | hi = i-1; |
| } while(lowOfs+1<hiOfs); | } |
| return not_found; | return not_found; |
| } | } |
| Line 382 static int transcodeFromUTF8(const XMLBy | Line 393 static int transcodeFromUTF8(const XMLBy |
| default: | default: |
| throw Exception(0, | throw Exception(0, |
| 0, | 0, |
| "transcodeFromUTF8 error: wrong trailingBytes value(%d)", trailingBytes); | "transcodeFromUTF8 error: wrong trailingBytes value(%d)", trailingBytes); // never |
| } | } |
| tmpVal-=gUTFOffsets[trailingBytes]; | tmpVal-=gUTFOffsets[trailingBytes]; |
| Line 392 static int transcodeFromUTF8(const XMLBy | Line 403 static int transcodeFromUTF8(const XMLBy |
| if(!(tmpVal & 0xFFFF0000)) { | if(!(tmpVal & 0xFFFF0000)) { |
| if(XMLByte xlat=xlatOneTo(tmpVal, tables, 0)) | if(XMLByte xlat=xlatOneTo(tmpVal, tables, 0)) |
| *outPtr++=xlat; | *outPtr++=xlat; |
| else | else { |
| outPtr+=sprintf((char *)outPtr, "&#%d;", tmpVal); // &#decimal; | outPtr+=sprintf((char *)outPtr, "&#%u;", tmpVal); // &#decimal; |
| } else | } |
| throw Exception(0, | } else { |
| 0, | const XMLByte* recoverPtr=srcPtr-trailingBytes-1; |
| "transcodeFromUTF8 error: too big tmpVal(0x%08X)", tmpVal); | for(uint i=0; i<=trailingBytes; i++) |
| outPtr+=sprintf((char*)outPtr, "%%%02X", *recoverPtr++); | |
| } | |
| } | } |
| // Update the bytes eaten | // Update the bytes eaten |
| Line 430 const String::C Charset::transcodeToUTF8 | Line 443 const String::C Charset::transcodeToUTF8 |
| (XMLByte *)src.str, src_length, | (XMLByte *)src.str, src_length, |
| dest_body, dest_length, | dest_body, dest_length, |
| tables)<0) | tables)<0) |
| throw(0, 0, | throw Exception(0, |
| 0, | 0, |
| "Charset::transcodeToUTF8 buffer overflow"); | "Charset::transcodeToUTF8 buffer overflow"); |
| assert(dest_length<=saved_dest_length); dest_body[dest_length]=0; // terminator | assert(dest_length<=saved_dest_length); dest_body[dest_length]=0; // terminator |
| return String::C((char*)dest_body, dest_length); | return String::C((char*)dest_body, dest_length); |
| } | } |
| static XMLCh change_case_UTF8(const XMLCh src, const Charset::UTF8CaseTable& table) { | |
| int lo = 0; | |
| int hi = table.size - 1; | |
| while(lo<=hi) { | |
| // Calc the mid point of the low and high offset. | |
| const unsigned int i = (lo + hi) / 2; | |
| XMLCh cur=table.records[i].from; | |
| if(src==cur) | |
| return table.records[i].to; | |
| if(src>cur) | |
| lo = i+1; | |
| else | |
| hi = i-1; | |
| } | |
| // not found | |
| return src; | |
| } | |
| static void store_UTF8(XMLCh src, XMLByte*& outPtr ) { | |
| if(!src) { | |
| // use the replacement character | |
| *outPtr++= '?'; | |
| return; | |
| } | |
| // Figure out how many bytes we need | |
| unsigned int encodedBytes; | |
| if(src<0x80) | |
| encodedBytes = 1; | |
| else if(src<0x800) | |
| encodedBytes = 2; | |
| else if(src<0x10000) | |
| encodedBytes = 3; | |
| else if(src<0x200000) | |
| encodedBytes = 4; | |
| else if(src<0x4000000) | |
| encodedBytes = 5; | |
| else if(src<= 0x7FFFFFFF) | |
| encodedBytes = 6; | |
| else { | |
| // use the replacement character | |
| *outPtr++= '?'; | |
| return; | |
| } | |
| // And spit out the bytes. We spit them out in reverse order | |
| // here, so bump up the output pointer and work down as we go. | |
| outPtr+= encodedBytes; | |
| switch(encodedBytes) { | |
| case 6: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL); | |
| src>>= 6; | |
| case 5: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL); | |
| src>>= 6; | |
| case 4: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL); | |
| src>>= 6; | |
| case 3: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL); | |
| src>>= 6; | |
| case 2: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL); | |
| src>>= 6; | |
| case 1: *--outPtr = XMLByte(src | gFirstByteMark[encodedBytes]); | |
| } | |
| // Add the encoded bytes back in again to indicate we've eaten them | |
| outPtr+= encodedBytes; | |
| } | |
| static void change_case_UTF8(XMLCh src, XMLByte*& outPtr, | |
| const Charset::UTF8CaseTable& table) { | |
| store_UTF8(change_case_UTF8(src, table), outPtr); | |
| }; | |
| void change_case_UTF8(const XMLByte* srcData, size_t srcLen, | |
| XMLByte* toFill, size_t toFillLen, | |
| const Charset::UTF8CaseTable& table) { | |
| const XMLByte* srcPtr=srcData; | |
| const XMLByte* srcEnd=srcData+srcLen; | |
| XMLByte* outPtr=toFill; | |
| XMLByte* outEnd=toFill+toFillLen; | |
| // We now loop until we either run out of input data, or room to store | |
| while ((srcPtr < srcEnd) && (outPtr < outEnd)) { | |
| // Get the next leading byte out | |
| const XMLByte firstByte =* srcPtr; | |
| if(firstByte<= 127) { | |
| change_case_UTF8(firstByte, outPtr, table); | |
| srcPtr++; | |
| continue; | |
| } | |
| // See how many trailing src bytes this sequence is going to require | |
| const unsigned int trailingBytes = gUTFBytes[firstByte]; | |
| // Looks ok, so lets build up the value | |
| uint tmpVal=0; | |
| switch(trailingBytes) { | |
| case 5: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 4: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 3: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 2: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 1: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 0: tmpVal+=*srcPtr++; | |
| break; | |
| default: | |
| throw Exception(0, | |
| 0, | |
| "change_case_UTF8 error: wrong trailingBytes value(%d)", trailingBytes); | |
| } | |
| tmpVal-=gUTFOffsets[trailingBytes]; | |
| // If it will fit into a single char, then put it in. Otherwise | |
| // fail [*encode it as a surrogate pair. If its not valid, use the | |
| // replacement char.*] | |
| if(!(tmpVal & 0xFFFF0000)) | |
| change_case_UTF8(tmpVal, outPtr, table); | |
| else | |
| throw Exception(0, | |
| 0, | |
| "change_case_UTF8 error: too big tmpVal(0x%08X)", tmpVal); | |
| } | |
| if(srcPtr!=outPtr) | |
| throw Exception(0, | |
| 0, | |
| "change_case_UTF8 error: end pointers do not match"); | |
| } | |
| const String::C Charset::transcodeFromUTF8(const String::C src) const { | const String::C Charset::transcodeFromUTF8(const String::C src) const { |
| size_t src_length=src.length; | size_t src_length=src.length; |
| size_t dest_length=src.length*6/*so that surly enough, "ÿ" has max ratio */; | size_t dest_length=src.length*6/*so that surly enough, "ÿ" has max ratio */; |
| Line 449 const String::C Charset::transcodeFromUT | Line 593 const String::C Charset::transcodeFromUT |
| (XMLByte *)src.str, src_length, | (XMLByte *)src.str, src_length, |
| dest_body, dest_length, | dest_body, dest_length, |
| tables)<0) | tables)<0) |
| throw(0, 0, | throw Exception(0, |
| 0, | 0, |
| "Charset::transcodeFromUTF8 buffer overflow"); | "Charset::transcodeFromUTF8 buffer overflow"); |
| Line 484 const String::C Charset::transcodeToChar | Line 628 const String::C Charset::transcodeToChar |
| static const Charset::Tables* tables[MAX_CHARSETS]; | static const Charset::Tables* tables[MAX_CHARSETS]; |
| #ifdef PA_PATCHED_LIBXML_BACKWARD | |
| #define declareXml256ioFuncs(i) \ | |
| static int xml256CharEncodingInputFunc##i( \ | |
| unsigned char *out, int *outlen, \ | |
| const unsigned char *in, int *inlen, void*) { \ | |
| return transcodeToUTF8( \ | |
| in, *(size_t*)inlen, \ | |
| out, *(size_t*)outlen, \ | |
| *tables[i]); \ | |
| } \ | |
| static int xml256CharEncodingOutputFunc##i( \ | |
| unsigned char *out, int *outlen, \ | |
| const unsigned char *in, int *inlen, void*) { \ | |
| return transcodeFromUTF8( \ | |
| in, *(size_t*)inlen, \ | |
| out, *(size_t*)outlen, \ | |
| *tables[i]); \ | |
| } | |
| #else | |
| #define declareXml256ioFuncs(i) \ | #define declareXml256ioFuncs(i) \ |
| static int xml256CharEncodingInputFunc##i( \ | static int xml256CharEncodingInputFunc##i( \ |
| unsigned char *out, int *outlen, \ | unsigned char *out, int *outlen, \ |
| Line 502 static const Charset::Tables* tables[MAX | Line 668 static const Charset::Tables* tables[MAX |
| *tables[i]); \ | *tables[i]); \ |
| } | } |
| #endif | |
| declareXml256ioFuncs(0) declareXml256ioFuncs(1) | declareXml256ioFuncs(0) declareXml256ioFuncs(1) |
| declareXml256ioFuncs(2) declareXml256ioFuncs(3) | declareXml256ioFuncs(2) declareXml256ioFuncs(3) |
| declareXml256ioFuncs(4) declareXml256ioFuncs(5) | declareXml256ioFuncs(4) declareXml256ioFuncs(5) |
| Line 531 void Charset::addEncoding(char *name_cst | Line 700 void Charset::addEncoding(char *name_cst |
| "already allocated %d handlers, no space for new encoding '%s'", | "already allocated %d handlers, no space for new encoding '%s'", |
| MAX_CHARSETS, name_cstr); | MAX_CHARSETS, name_cstr); |
| xmlCharEncodingHandler* handler=new(PointerFreeGC) xmlCharEncodingHandler; | xmlCharEncodingHandler* handler=new(UseGC) xmlCharEncodingHandler; |
| { | { |
| handler->name=name_cstr; | handler->name=name_cstr; |
| handler->input=inputFuncs[handlers_count]; | handler->input=inputFuncs[handlers_count]; |
| Line 544 void Charset::addEncoding(char *name_cst | Line 713 void Charset::addEncoding(char *name_cst |
| } | } |
| void Charset::initTranscoder(const StringBody NAME, const char* name_cstr) { | void Charset::initTranscoder(const String::Body NAME, const char* name_cstr) { |
| ftranscoder=xmlFindCharEncodingHandler(name_cstr); | ftranscoder=xmlFindCharEncodingHandler(name_cstr); |
| transcoder(NAME); // check right way | transcoder(NAME); // check right way |
| } | } |
| xmlCharEncodingHandler& Charset::transcoder(const StringBody NAME) { | xmlCharEncodingHandler& Charset::transcoder(const String::Body NAME) { |
| if(!ftranscoder) | if(!ftranscoder) |
| throw Exception("parser.runtime", | throw Exception("parser.runtime", |
| new String(NAME, String::L_TAINTED), | new String(NAME, String::L_TAINTED), |
| Line 562 String::C Charset::transcode_cstr(xmlCha | Line 731 String::C Charset::transcode_cstr(xmlCha |
| return String::C("", 0); | return String::C("", 0); |
| int inlen=strlen((const char*)s); | int inlen=strlen((const char*)s); |
| int outlen=inlen; // max | int outlen=inlen*6/*strlen("ÿ")*/; // max |
| #ifndef NDEBUG | #ifndef NDEBUG |
| int saved_outlen=outlen; | int saved_outlen=outlen; |
| #endif | #endif |
| Line 572 String::C Charset::transcode_cstr(xmlCha | Line 741 String::C Charset::transcode_cstr(xmlCha |
| if(xmlCharEncodingOutputFunc output=transcoder(FNAME).output) { | if(xmlCharEncodingOutputFunc output=transcoder(FNAME).output) { |
| error=output( | error=output( |
| (unsigned char*)out, &outlen, | (unsigned char*)out, &outlen, |
| (const unsigned char*)s, &inlen); | (const unsigned char*)s, &inlen |
| #ifdef PA_PATCHED_LIBXML_BACKWARD | |
| ,0 | |
| #endif | |
| ); | |
| } else { | } else { |
| memcpy(out, s, outlen=inlen); | memcpy(out, s, outlen=inlen); |
| error=0; | error=0; |
| Line 607 xmlChar* Charset::transcode_buf2xchar(co | Line 780 xmlChar* Charset::transcode_buf2xchar(co |
| int saved_outlen; | int saved_outlen; |
| #endif | #endif |
| if(xmlCharEncodingInputFunc input=transcoder(FNAME).input) { | if(xmlCharEncodingInputFunc input=transcoder(FNAME).input) { |
| outlen=buf_size*6/*max*/; | outlen=buf_size*6/*max UTF8 bytes per char*/; |
| #ifndef NDEBUG | #ifndef NDEBUG |
| saved_outlen=outlen; | saved_outlen=outlen; |
| #endif | #endif |
| out=(xmlChar*)xmlMalloc(outlen+1); | out=(xmlChar*)xmlMalloc(outlen+1); |
| error=input( | error=input( |
| out, &outlen, | out, &outlen, |
| (const unsigned char*)buf, (int*)&buf_size); | (const unsigned char*)buf, (int*)&buf_size |
| #ifdef PA_PATCHED_LIBXML_BACKWARD | |
| ,0 | |
| #endif | |
| ); | |
| } else { | } else { |
| outlen=buf_size; | outlen=buf_size; |
| #ifndef NDEBUG | #ifndef NDEBUG |
| Line 641 GdomeDOMString_auto_ptr Charset::transco | Line 818 GdomeDOMString_auto_ptr Charset::transco |
| return transcode_buf2dom(cstr, strlen(cstr)); | return transcode_buf2dom(cstr, strlen(cstr)); |
| } | } |
| GdomeDOMString_auto_ptr Charset::transcode(const StringBody s) { | GdomeDOMString_auto_ptr Charset::transcode(const String::Body s) { |
| const char* cstr=s.cstr(); | const char* cstr=s.cstr(); |
| return transcode_buf2dom(cstr, s.length()); | return transcode_buf2dom(cstr, s.length()); |
| } | } |
| #endif | #endif |
| String::Body Charset::transcode(const String::Body src, | |
| const Charset& source_transcoder, | |
| const Charset& dest_transcoder) { | |
| const char *src_ptr=src.cstr(); | |
| size_t src_size=strlen(src_ptr); | |
| String::C dest=Charset::transcode(String::C(src_ptr, src_size), | |
| source_transcoder, | |
| dest_transcoder); | |
| return String::Body(dest.str, dest.length); | |
| } | |
| String& Charset::transcode(const String& src, | |
| const Charset& source_transcoder, | |
| const Charset& dest_transcoder) { | |
| if(!src.length()) | |
| return *new String("", 0, false); | |
| return *new String(transcode((String::Body)src, source_transcoder, dest_transcoder), String::L_CLEAN); | |
| } | |
| void Charset::transcode(ArrayString& src, | |
| const Charset& source_transcoder, | |
| const Charset& dest_transcoder) { | |
| for(size_t i=0; i<src.count(); i++) | |
| src.put(i, &transcode(*src[i], source_transcoder, dest_transcoder)); | |
| } | |
| #ifndef DOXYGEN | |
| struct Transcode_pair_info { | |
| const Charset* source_transcoder; | |
| const Charset* dest_transcoder; | |
| }; | |
| #endif | |
| static void transcode_pair(const String::Body /*akey*/, | |
| String::Body& avalue, | |
| Transcode_pair_info* info) { | |
| avalue=Charset::transcode(avalue, | |
| *info->source_transcoder, | |
| *info->dest_transcoder); | |
| } | |
| void Charset::transcode(HashStringString& src, | |
| const Charset& source_transcoder, | |
| const Charset& dest_transcoder) { | |
| Transcode_pair_info info={&source_transcoder, &dest_transcoder}; | |
| src.for_each_ref(transcode_pair, &info); | |
| } |